of 22
Conformal SnO
x
Heterojunction Coatings
for
Stabilized Photoelectrochemical Water
Oxidation using Arrays
of Silicon
Microcones
Ivan A. Moreno-Hernandez
1
, Sisir Yalamanchili
3
, Harold J. Fu
1
, Harry A.
Atwater
3,4
, Bruce S.
Brunschwig
2
, Nathan S.
Lewis
1,2,4
1
Division
of Chemistry and Chemical
Engineering,
127-72, California Institute of Technology,
Pasadena, CA
91125, USA
2
Beckman Institute Molecular Materials
Research
Center, California Institute of Technology,
Pasadena, CA
91125, USA
3
Division of Engineering and Applied
Sciences,
California Institute of Technology, Pasadena,
CA
91125
4
Kavli Nanoscience Institute, California Institute of Technology, Pasadena,
CA 91125, USA
*Correspondence
to: nslewis@caltech.edu
S1
Electronic
Supplementary
Material
(ESI)
for
Journal
of
Materials
Chemistry
A.
This
journal
is
©
The
Royal
Society
of
Chemistry
2020
Supplementary Material
Materials and Methods
Materials Characterization
X-ray
diffraction
(XRD)
analysis
was
performed
with
a Bruker
D8
Discover
instrument
equipped
with
a Vantec-500
2-dimensional
detector.
Cu
radiation
(1.54
Å)
was
generated
at a
tube
current
of
1000
μA
and
a tube
voltage
of
50
kV.
The
incident
radiation
was
focused
with
a
0.5
mm
diameter
mono-capillary
collimator.
The
samples
were
placed
at
the
correct
position
for
diffraction
measurements
by
adjusting
the
sample
position
until
the
diffuse
reflectance
of
an
aligned
laser
beam
was
in
the
correct
position.
Measurements
were
collected
with
a coupled
theta/2theta
mode.
The
scattered
radiation
was
collected
with
a Vantec-500
detector
with
an
angular
resolution
of
< 0.04
°, which
enabled
the
collection
of
diffraction
from
a
range
of
20°
per
scan.
Four
scans
were
collected
in
the
range
of
20°
to 80°
2θ,
and
radiation
was
counted
for
a
total
duration
of
4 h to
obtain
the
XRD
profile.
The
collected
data
were
analyzed
using
Bruker
EVA
software.
The
diffraction
data
were
compared
to reference
patterns
for
SnO
2
.
1
An
FEI
Tecnai
F30ST
transmission-electron
microscope
with
an accelerating
voltage
of
300
kV
was
used
to
image
the
n-Si/SiO
x
/SnO
x
interface.
A
cross
section
was
prepared
via
mechanical
polishing
and
ion-milling
techniques,
which
limited
the
TEM
analysis
to planar
Si substrates.
Scanning-electron
microscope
images
were
collected
with
a
Nova
nanoSEM
450
(FEI)
instrument on both planar
and microcone array
photoelectrodes.
Surface Recombination Velocity Measurements
Carrier
lifetimes
were
determined
with
an
instrument
and
procedure
described
in
previous
work.
2
Intrinsically
doped
300
μm
thick
silicon
wafers
polished
in
the
(111)
orientation
S2
with
a bulk
lifetime
> 1.5
ms
were
chemically
oxidized
to
form
a controlled
SiO
x
layer
and
coated
with
100
SnO
x
ALD
cycles
deposited
at 210
°C
through
the
same
procedure
that
was
used
for
n-type
silicon
wafers.
Intrinsically
doped
Si
wafers
were
utilized
for
this
experiment
because
the
high
lifetime
of
intrinsically
doped
Si
allows
the
observed
recombination
to
be
attributed
to
surface
recombination.
Electron-hole
pairs
were
generated
in the
i-Si/SiO
x
/SnO
x
sample
by
a 20
ns,
905
nm
laser
pulse
from
an
OSRAM
diode
laser
with
an ETX-10A-93
driver.
A
PIN
diode
connected
to
an
oscilloscope
was
used
to
measure
the
decay
in reflected
microwave
intensity
after
each
excitation.
The
lifetime
was
determined
by
fitting
the
average
decay
curve
of
64
consecutive
scans
with
an
exponential
function.
The
surface
recombination
velocity
was
determined from the following relationship:
1
=
1
+
2푆
2푆
where
S
is the surface
recombination velocity,
d
is the thickness of the silicon
wafer,
τ
m
is the
bulk lifetime, and
τ
m
is the
measured
lifetime
of the electron-hole pairs.
X-ray Photoelectron Spectroscopy
X-ray
photoelectron
spectroscopic
(XPS)
data
were
collected
using
a Kratos
Axis
NOVA
(Katros
Analytical,
Manchester,
UK)
at a background
pressure
of
< 10
-9
Torr.
A monochromatic
Al
source
at
1486.6
eV
was
used
for
excitation.
High-resolution
scans
were
collected
at
a
resolution
of
0.05
eV
and
survey
scans
were
collected
at 1 eV
resolution.
All
energies
including
peak
energies,
valence-band
spectra,
and
work-function
measurements
were
calibrated
against
the
binding
energy
of the
adventitious
C 1s
peak,
which
was
set
at 284.8
eV.
For
core-level
peak
energies
and
valence-band
spectra,
the
bias
of
the
sample
was
controlled
by
the
XPS
instrument,
S3
whereas
for
work-function
measurements
the
bias
of
the
sample
was
controlled
by
an external
power
supply.
Although
this
study
characterized
both
planar
and
microcone
arrays
with
photoelectrochemical
methods,
XPS
data
were
only
collected
for
planar
electrodes.
The
acquired
XPS
signal
has
a strong
dependence
on
the
angle
between
the
electron
emission
path
and
sample
surface,
so
data
collection
on microcone
arrays
would
result
in convolution
of
the
acquired
signal
with the sample geometry, complicating the quantitative analysis
of XPS data.
Using
a previously
reported
procedure,
the
Ni
2p
spectra
were
fit
to determine
the
composition
of
the
electrodeposited
NiFeOOH
film.
3
Due
to the
small
projected
area
coverage
of
the
catalyst
on
the
n-Si/SiO
x
/SnO
x
electrode
in
conjunction
with
the
overlap
of
other
elemental
peaks,
the
Fe
2p
spectra
could
not
be
used
to
quantify
the
Fe
in
the
electrodeposited
film.
The
Ni
2p
3/2
spectra
of
n-Si/SiO
x
/SnO
x
/e-NiFeOOH
electrodes
that
had
been
subjected
to
3-5
cyclic
voltammetric
scans
were
fit
with
the
reported
peak
separations,
FWHM
ratios,
and
relative
peak
areas
for
standard
Ni
samples.
To
account
for
the
precision
of
XPS
measurements,
the
absolute
peak
positions
were
allowed
to
vary
within
±0.1
eV
of
the
reported
peak
position,
but
the
reported
peak
separations
were
not
adjusted.
Only
contributions
from
Ni(OH)
2
were
required
to
adequately fit the collected XPS
data for
the electrodeposited films.
The
Co
2p
spectra
were
fit
with
a previously
reported
procedure
to
determine
the
amount
of
Co
3
O
4
and
Co(OH)
2
in
the
electrodeposited
CoO
x
film.
4
The
Co
2p
3/2
spectra
of
n-
Si/SiO
x
/SnO
x
/CoO
x
electrodes
operated
for
3-5
cyclic
voltammetric
scans
were
fit
with
the
reported
peak
separations,
FWHM
ratios,
and
relative
peak
areas
for
standard
Co
samples.
To
account
for
the
precision
of
XPS
measurements,
the
absolute
peak
positions
were
allowed
to
vary
within
±0.1
eV
of
the
reported
peak
position,
but
the
reported
peak
separations
were
not
adjusted.
Contributions
from
both
Co
3
O
4
and
Co(OH)
2
were
required
to adequately
fit
the
S4
collected
XPS
data
for
the
electrodeposited
films.
The
Ir
4f
spectra
were
fit
based
on
the
previously
reported
XP
spectra
of Ir
electrodes
for
water
oxidation
in sulfuric
acid.
5
Sn
3d
spectra
could
be
adequately
described
by
a single
component,
and
this
was
assigned
to
SnO
2
based
on the
previously
reported
binding
energy
ranges.
6
Si
2p
spectra
were
fit
with
previously
reported
peak
shapes.
7
All
x-ray
photoelectron
spectra
were
processed
using
CasaXPS.
A Shirley
background
was
utilized
to
correct
for
the
background
XPS
signal.
Most
peaks
were
fit
to
Gaussian/Lorentzian
symmetric
line-shapes
with
a 3:7
Gaussian:Lorentzian
contribution
unless
specified by the literature reference used to assign the XPS peaks.
Electrochemical Testing
The
1.0
M
KOH(aq)
and
1.0
M
H
2
SO
4
(aq)
electrolytes
were
prepared
by
dissolving
or
diluting
potassium
hydroxide
pellets
(Sigma-Aldrich,
Semiconductor
Grade,
99.99%)
and
sulfuric
acid
(Fischer
Scientific,
TraceMetal
Grade,
93-98%)
with
deionized
water
(18.2
MΩ
cm),
respectively.
The
Fe(CN)
6
3-/4-
(aq)
electrolyte
was
prepared
by
dissolving
potassium
ferrocyanide
trihydrate,
potassium
ferricyanide,
and
potassium
chloride
in
deionized
H
2
O
to
make
a 0.36
M
K
4
Fe(CN)
6
/0.05
M
K
3
Fe(CN)
6
/0.50
M
KCl
solution.
Digital
potentiostats
were
used
to
collect
the
electrochemical
data
(SP-200
or
MPG-2,
Bio-Logic
Science
Instruments).
The
reference
electrodes
used
in
1.0
M
KOH(aq),
1.0
M
H
2
SO
4
(aq),
and
Fe(CN)
6
3-/4-
(aq)
were
mercury/mercury
oxide
(Hg/HgO,
CH
Instruments),
mercury/mercury
chloride
(SCE,
CH
Instruments),
and
a Pt disk
(CH
instruments),
respectively.
The
Hg/HgO
and
SCE
reference
electrodes
were
calibrated
versus
a reversible
hydrogen
electrode
(RHE)
in
1.0
M KOH(aq)
and
1.0
M
H
2
SO
4
(aq),
respectively.
The
RHE
consisted
of a Pt
disk
in H
2
-saturated
electrolyte,
with
constant
bubbling
of
H
2
(g)
underneath
the
Pt disk
to
ensure
H
2
saturation.
The
calibrated
potentials
of
the
Hg/HgO
and
SCE
were
0.906
V
and
0.
244
V
vs.
RHE,
respectively.
The
S5
counter
electrode
was
a
carbon
rod
in
1.0
M
KOH(aq)
or
1.0
M H
2
SO
4
(aq).
Cyclic
voltammograms
with
Pt working,
reference,
and
counter
electrodes
were
used
to verify
that
the
Pt
equilibrated
to
the
Fe(CN)
6
3-/4-
(aq)
Nernstian
cell
potential.
The
cyclic
voltammograms
exhibited
zero
current
at zero
voltage
versus
the
Pt
reference
electrode.
The
Nernstian
potential
for
Fe(CN)
6
3-/4-
(aq)
was
0.31
V
vs.
SHE
based
on
the
concentrations
of
the
redox
species.
All
glassware
was
cleaned
with
aqua
regia
and
rinsed
with
deionized
water
prior
to
use.
The
electrolyte
volume
was
generally
~ 40
mL
and
cyclic
voltammetric
data
were
collected
at
a scan
rate of 40 mV s
-1
.
Impedance
measurements
in
contact
with
Fe(CN)
6
3-/4-
(aq)
were
collected
without
illumination
over
a frequency
range
of
20
Hz
to
20
kHz
with
a sinusoidal
wave
amplitude
of
10
mV.
A potential
of
-0.3
to
0.3
V vs.
Fe(CN)
6
3-/4-
was
applied
to p
+
-Si/SiO
x
/SnO
x
electrodes
and
a
potential
of 0.0
to
0.6
V
vs.
Fe(CN)
6
3-/4-
was
applied
to n-Si/SiO
x
/SnO
x
electrodes.
The
impedance
measurements
were
fit
with
a circuit
consisting
of
a resistor
in
series
with
an
additional
component
consisting
of
a resistor
and
a capacitor
in parallel,
as
has
been
described
previously.
8
The
differential
capacitance
determined
from
the
impedance
measurements
on
n-
Si/SiO
x
/SnO
x
electrodes
was
attributed
to
the
differential
capacitance
of
the
depletion-region
of
the
silicon
semiconductor.
The
differential
capacitance
determined
from
the
impedance
measurements
on
p
+
-Si/SiO
x
/SnO
x
electrodes
was
attributed
to the
SnO
x
layer,
due
to the
voltage
dependence
suggesting
n-type
behavior.
The
potential
dependence
of
the
differential
capacitance
was analyzed
using the Mott-Schottky relationship:
2
=
2
2
0
(
푎푝푝
+
푏푖
)
S6
where
C
d
is the
differential
capacitance,
q
is
the
unsigned
charge
of
an
electron,
A
is
the
electrode
area,
ε
0
is the
vacuum
permittivity,
ε
r
is the
relative
permittivity
of
the
semiconductor,
N
d
is
the
donor
dopant
concentration
of
the
semiconductor,
V
app
is
the
electrode
potential
vs.
Fe(CN)
6
3-/4-
(aq),
V
bi
is the
built-in
voltage
in
the
semiconductor,
and
T
is the
temperature
of
the
electrode
while
the
impedance
data
were
collected.
A
relative
permittivity
of
ε
r
= 11.7
was
used
to
analyze
data
from
n-Si/SiO
x
/SnO
x
electrodes,
consistent
with
previous
reports
for
Si
photoanodes.
A
relative
permittivity
of
ε
r
= 11.5
was
used
to
analyze
data
from
p
+
-Si/SiO
x
/SnO
x
electrodes,
corresponding
to
the
average
relative
permittivity
of
SnO
2
.
9
The
donor
dopant
concentration
N
d
was
attributed
to
the
n-Si
semiconductor
for
n-Si/SiO
x
/SnO
x
electrodes,
and
to
SnO
x
for p
+
-Si/SiO
x
/SnO
x
electrodes.
The
dissolution
of
SnO
x
deposited
with
850
cycles
at 210
°C
on
p
+
-Si/SiO
x
under
η
= 300
mV
for
water
oxidation
was
determined
in
1.0
M KOH(aq)
and
1.0
M H
2
SO
4
(aq).
Electrodes
with
p
+
-Si
instead
of
n-Si
were
used
to
precisely
set
the
surface
potential
of
SnO
x
.
Utilizing
an n-
Si
substrate
instead
of p
+
-Si
would
lead
to
convolution
with
the
photovoltage
generated
under
illumination,
which
can
vary
due
to
light
intensity
variability
throughout
the
> 500
h dissolution
test.
Cells
for
these
experiments
had
working
and
reference
electrodes
separated
from
the
counter
electrode
by
an
ion-exchange
membrane.
An
anion-exchange
membrane
(Fumasep
FAA-3-PK-
130)
and
a cation-exchange
membrane
(Nafion
117)
were
used
for
1.0
M KOH(aq)
and
1.0
M
H
2
SO
4
(aq)
electrolytes,
respectively.
Electrodes
with
an area
of
~ 0.1
cm
2
were
held
at
η
= 300
mV
in an
initial
50
mL
volume
of electrolyte.
Samples
of
200
uL
in volume
were
taken
over
time
and
diluted
to
5 mL
with
deionized
water.
The
concentration
of
Sn
was
determined
using
calibration
solutions
(Semi
Metals
Plasma
Standard
Solution,
Specpure,
Alfa
Aesar)
and
an
inductively
coupled
plasma
mass
spectrometer
(ICP-MS,
Agilent
800
Triple
Quadrupole).
The
S7
concentration
of
Sn
was
used
to
determine
the
equivalent
amount
of
Sn
lost
from
the
electrode,
assuming
a stoichiometry
of SnO
2
and
a density
equal
to the
bulk
density
of
SnO
2
(6.95
g cm
-3
).
The
amount
of
Sn
removed
from
the
electrolyte
during
the
sample
acquisition
(200
uL)
was
taken into account when
determining total amount of
Sn
dissolved from the electrode over
time.
Electrochemical Depositions
Nickel-iron
oxyhydroxide
(NiFeOOH)
was
electrochemically
deposited
without
illumination
at a cathodic
current
density
of
-1
mA
cm
-2
on
planar
electrodes
and
-10
mA
cm
-2
on
microcone
arrays.
The
stirred
aqueous
solution
contained
5×10
-3
M
Ni(NO
3
)
2
and
5×10
-3
M
FeSO
4
in
a two-electrode
experiment
with
a single-compartment
electrochemical
cell
that
had
a
carbon
rod
as
a counter
electrode.
The
Ni(NO
3
)
2
(aq)
solution
was
purged
with
N
2
(g)
prior
to the
addition
of FeSO
4
6H
2
O(s)
to
prevent
Fe
precipitation.
Cobalt
oxide
(CoO
x
)
was
photoelectrochemically
deposited
from
a 0.01
M
Co(II)
nitrate/0.10
M
sodium
acetate
aqueous
solution
under
100
mW
cm
-2
of
simulated
solar
illumination
with
an anodic
current
density
of
1
mA
cm
-2
,
using
a two-electrode
experiment
with
a single-compartment
electrochemical
cell
with
a carbon
rod
as a counter
electrode.
The
NiFeOOH
and
CoO
x
loadings
were
controlled
by
changing
the
duration
of
the
chronopotentiometric
deposition.
Iridium
oxide
(IrO
x
)
was
photoelectrochemically
deposited
using
a modified
literature
procedure.
10
Aqueous
dark-brown
solutions
of
2×10
-3
M
K
2
IrCl
6
and
2×10
-3
M
KOH
were
heated
at 70
°C
under
continuous
stirring
until
the
solution
became
light-brown
and
slightly
turbid.
The
solution
was
then
placed
in an
ice
bath
to
prevent
precipitation
of
Ir
species.
The
electrochemical
cell
consisted
of
one
compartment
with
a SCE
as a reference
electrode
and
a carbon
as
a counter
electrode.
IrO
x
was
deposited
by
using
cyclic
voltammetric
scans
from
-0.2
to 1.2
V
vs.
SCE
under
100
mW
cm
-2
of
S8
simulated
solar
illumination.
The
loading
of
the
IrO
x
was
controlled
by
changing
the
number
of
cyclic voltammetric scans.
Photovoltage Dependence
on Surface
Area
For
junctions
that
have
a constant
surface-area-normalized
diode
current
density
that
is
independent
of
the
geometry
of
the
structure,
as
is
expected
for
well-formed
junctions,
the
observed photovoltage scales with the surface area
increase,
γ,
based on equation 1:
11
(1)
표푐
=
ln
(
푝ℎ
0
)
For
the
SnO
x
heterojunction,
the
ideality
factor
was
determined
to be
n
= 1.07
and
the
diode
current
density
was
J
0
= 1.3*10
-11
A
cm
-2
for
photocurrents
between
15
– 60
mA
cm
-2
.
Planar
samples
exhibited
γ
= 1,
which
results
in
a photovoltage
of
596
mV
at
a photocurrent
of
30
mA
cm-
2
,
close
to
that
observed
in
Figure
S1a.
The
n-Si
microcone
arrays
studied
herein
had
a pitch
of
7
μm,
a height
of
~60
μm,
and
a base
diameter
of
3.5
μm.
The
surface
area
calculated
from
these
dimensions
leads
to
γ
~
13.3.
The
photovoltage
expected
for
n-Si
microcone
array
electrodes
based
on
equation
1 is
thus
524
mV,
for
γ
= 13.3
at
30
mA
cm
-2
of
photocurrent
density, which
is close to the experimentally
observed photovoltage of 490 mV (Figure S9a).
S9
Supplementary Information Figures
Figure
S1.
Electrochemical
behavior
of
n-Si/SiO
x
/SnO
x
photoanodes
coated
with
SnO
x
deposited
at
temperatures
between
150–250
°C
in contact
with
Fe(CN)
6
3-/4-
(aq).
a)
Photoelectrochemical
behavior
of n-Si/SiO
x
/SnO
x
photoanodes
prepared
with
100
SnO
x
ALD
cycles
under
100
mW
cm
-2
of
simulated
solar
illumination,
and
n-Si/SnO
x
photoanodes
prepared
by
depositing
SnO
x
on
HF-etched
n-Si
substrates.
b)
Open-circuit
voltage
of
n-Si/SiO
x
/SnO
x
S10
photoanodes
prepared
with
100
or
500
SnO
x
ALD
cycles
under
100
mW
cm
-2
of simulated
solar
illumination.
c)
Mott-Schottky
plot
of
electrode
differential
capacitance
versus
voltage
of n-
Si/SiO
x
/SnO
x
photoanodes
in
the
dark
with
a SnO
x
layer
that
was
deposited
at
210
°C
with
100
ALD
cycles.
d)
Open-circuit
voltage
versus
short-circuit
current
density
(
V
oc
vs
J
sc
)
plot
of
-
Si/SiO
x
/SnO
x
deposited
at
210
°C
in
contact
with
Fe(CN)
6
3-/4-
(aq)
under
1 – 200
mW
cm
-2
of
simulated solar illumination.
Figure
S2.
a)
Dark
field
image
of
n-Si(100)/SiO
x
coated
with
850
cycles
of
SnO
x
with
image
brightness
corresponding
to
(002)
SnO
2
diffraction
signal
collected
through
a 10
μm
diemeter
aperture.
b) Electron
diffraction
pattern
of
n-Si/SiO
x
coated
with
850
cycles
of
SnO
x
,
at
300
kV
accelerating voltage.
S11
Figure
S3.
X-ray
diffraction
pattern
for
a SnO
x
film
deposited
at
210
°C
with
850
ALD
cycles
on an n-Si(100)/SiO
x
substrate. Weak diffraction signals
associated with SnO
2
were observed.
Figure
S4.
Dissolution
of
SnO
x
film
deposited
at
210
°C
with
850
ALD
cycles
on
a p
+
-
Si(100)/SiO
x
substrate.
Electrodes
were
held
at
1.53
V vs.
RHE
in 1.0
M KOH(aq)
or
1.0
M
H
2
SO
4
(aq) eletrolytes.
S12
Figure
S5.
Optical
reflectivity
measurements
at
normal
incidence
of
an
n-Si/SiO
x
/SnO
x
wafer
coated
with
SnO
x
formed
at
210
°C
with
850
ALD
cycles
compared
to an n-Si
wafer
with
a
native silicon oxide layer.
S13
Figure
S6.
Electrochemical
behavior
of
n-Si/SiO
x
/SnO
x
photoanodes
coated
with
SnO
x
deposited
at
210
°C
in
contact
with
Fe(CN)
6
3-/4-
(aq)
under
100
mW
cm
-2
of
simulated
solar
illumination.
a)
Photoelectrochemical
behavior
of
n-Si/SiO
x
/SnO
x
photoanodes
with
100
or
850
ALD
SnO
x
cycles.
b)
Photoelectrochemical
behavior
of
n-Si/SiO
x
photoanodes
with
850
cycles
of
SnO
x
with
and
without
a sputtered
Pt
overlayer.
c)
Mott-Schottky
plot
of
electrode
differential
apacitance
versus
potential
of
p
+
-Si/SiO
x
/SnO
x
electrodes
in the
dark
with
a SnO
x
layer
deposited
S14
at
210
°C
with
850
ALD
cycles.
d)
Mott-Schottky
plot
of
electrode
differential
capacitance
versus
potential
for p
+
-Si/SiO
x
/SnO
x
/Pt
in the
dark, indicating no apparent
band bending.
Figure
S7.
Photoelectrochemical
behavior
of
planar
n-Si/SiO
x
photoanodes
coated
with
100
cycles
of
SnO
x
and
electrodeposited
water
oxidation
catalysts
in contact
with
1.0
M
KOH(aq)
under
100
mW
cm
-2
of simulated
solar
illumination.
a)
Photoanodes
with
different
amounts
of
NiFeOOH
electrodeposited
in
the
dark
at
a current
density
of
1 mA
cm
-2
.
b)
Photoanodes
with
different
amounts
of
CoO
x
photoelectrodeposited
under
100
mW
cm
-2
of
simulated
solar
illumination
at
a photocurrent
density
of 1 mA
cm
-2
.
c)
Photoanodes
with
different
amounts
of
IrO
x
photoelectrodeposited
under
100
mW
cm
-2
of
simulated
solar
illumination
with
a cyclic
voltammetry procedure.
S15